About Me

True story: I studied software engineering because I wanted to make games. Somewhere during those years at school (in the early to mid 90s), after researching the game development business, I decided that maybe the work-life balance of game development wasn't for me and I pivoted to general software engineering.

And now, after many years of different roles in the development of (re)insurance systems, I have an opportunity to play game designer as a hobby, without it taking up all my spare time. Which is pretty friggin' great. Those who know me know that I love my hobbies to be something I can think about in my spare time — while commuting, walking, or other quiet moments. It's one of the reasons that I'm rarely, if ever, bored. Thinking about new game designs or how to improve the ones I've started works perfectly for this. As long as consumer AI remains affordable, I hope I can continue using it to make games.

How I Got Started

I ended up buying a year of Claude as an impulse purchase after seeing the news in February 2026 about Anthropic rejecting illegal US government use of their tools for mass surveillance and autonomous killing. That, and the subsequent news the very next day that Sam Altman swooped in and agreed to let OpenAI fill that void angered me to the point where I deleted my OpenAI account (which I had only barely used while experimenting with ChatGPT) and bought my year of Claude. Next was a bit over a month of "now that I have Claude, what should I use it for?". Eventually, I heard enough about how impressive its coding abilities were to force myself to try it out.

I honestly don't remember how I set up the various components on my home laptop, but I can tell you how I did it in case you're thinking of doing the same: I just asked Claude to walk me through it step by step.

There were two key prompts. One to start work on the design doc and another to help walk me through the installation and configuration I'd need to do. I found the first but somehow lost the second.

Here's the first:

"Hi Claude. I'd like to begin work on a design document for a game that I will build in Claude Code. In this chat I'll add details until I'm fairly sure I've incorporated all the major details, then I'll ask you to ask me any clarifying questions you'd need to complete that document and ensure it has all that's needed for successful Claude Code generation. This will be my first project using Claude Code. Quick check — does this process make sense? Anything I should know before I start?"

And I just asked Claude to generate a prompt for someone who wants to get started doing the same. Here it is:

"I'm a complete beginner who wants to make browser-based games using HTML5, JavaScript, and Canvas (or a lightweight game library). Please give me step-by-step instructions to set up my development environment from scratch on [Windows/Mac]. Include:

• What software to install (e.g. VS Code, Node.js, Git/GitHub Desktop) with links and recommended settings
• Any helpful VS Code extensions for game development
• Whether to use a game library like Phaser or plain canvas, with a brief reason why
• How to set up a basic project folder structure
• How to run and preview my game locally in a browser
• How to put my project on GitHub
• How to choose an option to host it at no cost

Assume I have no prior coding experience and explain each step clearly."

The beauty of it is also that if you get stuck (and I did several times), you just feed Claude chat back the errors you're seeing (i.e. both via copy/paste AND screenshots) and/or the things you don't understand. Claude will hold your hand and patiently walk you through it.

This is actually a pattern I've grown to use all the time now. If I don't know how to do something with Claude or see any issues, I just ask it for advice on how to resolve it. It's been a difficult habit to start, due to 35+ years of "muscle memory" where I manually search for info and try to come up with an approach myself.

I'm ambivalent about this: One one hand, its so useful and easy -- the quality of my answers from AI has been typically excellent. But on the other, I have to remember to use my "struggle a little and figure it out muscles" to make sure I practice and maintain my problem-solving skills. Honestly, leaning on AI is a bit too easy!